ZACH Insights / The Achilles tendon
Three muscles, one tendon. And it twists.
The Achilles tendon carries the pull of the calf to the heel bone. It is built from three separate strands that rotate around each other on the way down, it stiffens when it is loaded hard for months, and its core is hardly renewed in adult life. This article follows the tendon from the calf to the heel and shows what the research can and cannot say.
By Zacharias Razvi · Updated 14 November 2026 · 13 min read
From calf to heel
A cable between the calf and the heel.
The Achilles tendon, in anatomy books called the tendo calcaneus, runs from the lower half of the calf to the back of the heel bone, the calcaneus. Three muscles pull on it: the two heads of the gastrocnemius, which start above the knee, and the soleus, which lies underneath and starts on the shin bones. Together they are called the triceps surae, the three-headed muscle of the calf. 1
Every time the heel lifts in walking, running or jumping, the force goes through this tendon. Direct measurements in volunteers, with a sensor fitted on the tendon, found forces that were unexpectedly high in hopping. The authors suggested that how fast the load rises may matter more clinically than how large it is. 2
The tendon is also elastic. It stretches a little under load and recoils, and how much it stretches for a given force can change with training. Before that, it helps to see what it is made of.
General education. The article explains anatomy and research. It does not diagnose and it is not individual advice.
Model 1 · The calf from behind
Three muscles feed one tendon.
A right calf seen from behind, traced from a classic anatomy plate. Select a step to follow each muscle into the tendon.
-
One tendon, three sources.
The two heads of the gastrocnemius (1) and the soleus (2) all end in the Achilles tendon (4). The small plantaris (3) runs alongside it. The tendon inserts on the back of the calcaneus.
The plate comes from the 1918 edition of Gray's Anatomy, which is in the public domain. The gastrocnemius is cut away in the plate so that the soleus can be seen.
-
Two heads that cross the knee.
The medial and lateral heads of the gastrocnemius start on the back of the thigh bone, just above the knee. Their bellies form the visible bulge of the calf and end about halfway down the leg in a broad sheet of tendon that joins the Achilles tendon. 1
Because they cross the knee, the gastrocnemius pulls hardest when the knee is straight. In the plate both heads are cut through below the knee. Their path down to the tendon is shown with a dashed line.
-
A flat muscle underneath.
The soleus lies under the gastrocnemius and starts on the back of the tibia and the fibula. It does not cross the knee, so it can pull on the tendon whatever the knee is doing. 1
Its fibres reach the tendon lower down than those of the gastrocnemius. In the plate, the lowest soleus fibres on the outer side reach almost to the ankle.
-
A thin extra tendon.
The plantaris is a small muscle at the back of the knee with a very long, thin tendon. In the plate it runs down the inner side of the calf, between the gastrocnemius and the soleus, and continues along the inner edge of the Achilles tendon towards the heel. 1
It is too small to add much force. It is shown here because it lies close to the Achilles tendon and can be mistaken for part of it.
-
The tendon and its insertion.
Below the muscles the tendon runs free for a few centimetres and then inserts on the back of the calcaneus. In 106 legs from 53 male body donors, the distance from the lowest muscle fibres to the top of the heel bone was 60.8 mm on average, with a standard deviation of 14.2 mm. 4
The same study found that the tendon is thinner where the muscles end than where it attaches to the bone. The next model looks inside it.
Inside the tendon
The strands keep their identity.
The Achilles tendon is not one uniform band. Dissections show that it is made of three bundles of fibres, often called subtendons, one from each of the three muscle parts. The bundles stay separate along the tendon and rotate around each other on the way to the bone. 3 4
- Who took part
- 53 male body donors, 106 legs, in a study by Pękala and colleagues from Kraków. 4
- What was done
- Each subtendon was separated from the others and its rotation from the muscle to the bone was measured with a purpose-built 3D-printed tool.
- What they found
- Fibres from the medial head of the gastrocnemius rotated 28° on average. Fibres from the lateral head rotated 136° and fibres from the soleus 129°. At the insertion, the lateral head made up 44% of the tendon, the soleus 28% and the medial head 28%.
- What it cannot say
- How the twist behaves in a living, loaded tendon, and whether it differs between women and men. All donors were men, and the measurements were made on dissected tissue.
Model 2 · The twist
Follow each strand to the bone.
The tendon on the left turns slowly so that the strands can be seen from all sides. Move the level to see the cross-section at that height.
-
Three strands, three amounts of turn.
The strands start side by side below the muscles and end in new positions at the bone. They turn by very different amounts. Select a strand to follow it alone.
-
Medial gastrocnemius: 28°.
The strand from the inner head starts at the back and inner side and stays near the back of the tendon. On average it turns less than a tenth of a full turn. 4 Edama and colleagues found that it ran fairly straight in every tendon they examined. 3
-
Lateral gastrocnemius: 136°.
The strand from the outer head starts at the back and outer side and turns more than a third of a turn. 4 In three of four tendons in a Japanese study, it ended in the deep part of the tendon, nearest the ankle joint. 5
-
Soleus: 129°.
The soleus strand starts in front, in the deep layer, and turns almost as far as the lateral head. 4 Edama and colleagues found that most of its rotation happens 3–5 cm above the insertion. 3
-
The shares at the bone.
Where the tendon meets the calcaneus, the lateral head makes up the largest part, 44% of the cross-section. The soleus and the medial head make up 28% each. 4
The shares and angles are averages. The standard deviation of the angles was 15° for the medial head and 30–34° for the other two, so single tendons differ a good deal.
Variation between people
Three types of twist.
The amount of twist is not the same in everyone. A Japanese group sorted tendons into three types by which strand fills the deep layer at the insertion, the part nearest the ankle joint. 5 In the least twisted type, only the soleus lies there. In the moderate type, the soleus and the lateral head share it. In the most twisted type, only the lateral head lies there.
- Who took part
- 132 legs from 74 body donors, mean age at death 78 years, 87 legs from men and 45 from women. 5
- What was done
- The tendon was dissected down to the calcaneus, and the strands found in the deep layer of the insertion were recorded.
- What they found
- Least twist in 31 legs (24%), moderate in 87 (67%) and extreme in 12 (9%). The proportions were the same in men and women.
- What it cannot say
- Whether the type affects the risk of injury. No one has followed people with known twist types to see who is injured. The authors suggest that the types may load the deep fibres differently, which is a hypothesis.
Data · Twist types at the insertion
Which strand fills the deep layer, and how often.
Type I, soleus deep: 31 legs, 24%. Type II, soleus and lateral head deep: 87 legs, 67%. Type III, lateral head deep: 12 legs, 9%.
What training changes
Heavy loading makes the tendon stiffer.
A tendon stretches when it is pulled. Stiffness describes how much force it takes to stretch it by a given length. A stiffer tendon stretches less under the same force. Two properties decide stiffness: the material the tendon is made of, measured as Young's modulus, and how thick the tendon is, measured as its cross-sectional area.
Many training studies have measured these properties in the Achilles and patellar tendons with ultrasound. Two meta-analyses have pooled them. 6 7
Model 3 · Force and stretch
What a stiffer tendon means.
Choose a force and compare the stretch before and after a period of heavy loading.
At the same force, the tendon after training stretches less than before.
- Who took part
- People without tendon problems. Bohm and colleagues pooled 27 studies with 264 participants aged 18 to 50. 6 Lazarczuk and colleagues pooled 61 studies with 763 participants. 7
- What was compared
- Tendon properties before and after training programmes, most of them resistance training. Bohm included only programmes of at least 8 weeks.
- What they found
- Stiffness and Young's modulus rose by moderate to large amounts. Thickness rose only a little. In both reviews the change in material, not the change in thickness, explained most of the gain in stiffness. Bohm found larger gains with high loads and with programmes of 12 weeks or more.
- What it cannot say
- Whether a stiffer tendon is less likely to be injured. The studies measured healthy tendons for weeks to months, not injuries.
Data · Two meta-analyses
Change after training, as standardised mean differences.
Stiffness 0.70 (0.51 to 0.88) and 0.74 (0.62 to 0.86). Young's modulus 0.69 (0.36 to 1.03) and 0.82 (0.58 to 1.07). Cross-sectional area 0.24 (0.07 to 0.42) and 0.22 (0.12 to 0.33).
A clock from the 1960s
The core of the tendon is hardly renewed.
If a tendon adapts, does it replace its tissue? A Danish group answered this with a method that uses the nuclear bomb tests of 1955 to 1963. The tests raised the amount of carbon-14 in the air sharply, and the level has fallen since. Plants and animals take up carbon from the air, so tissue built in a given year carries that year's carbon-14 level. Tissue that is renewed keeps catching up with the falling level. Tissue that is not renewed keeps the level of the years it was built. 8
- Who took part
- Tissue from deceased people born between 1945 and 1983: 28 samples from the core of the Achilles tendon and 4 samples of muscle. 8
- What was done
- Carbon-14 in each sample was measured and compared with the known levels in the air over the donor's lifetime.
- What they found
- The tendon core matched the air during roughly the first 17 years of life, the years of growth in height. Muscle matched recent air, as expected from tissue that is renewed all the time.
- What it cannot say
- What happens on the surface of the tendon and in its sheath, which were not measured. A tendon can change its stiffness without new core tissue, as the previous section showed.
Model 4 · Reading the bomb pulse
Pick a birth year.
The tendon core keeps the carbon-14 level of the years of growth. Muscle keeps up with the air. Move the birth year to see how far apart they end up.
For someone born in 1950, the tendon core carries the high carbon-14 level of the 1950s and 1960s, while muscle carries today's low level.
Injuries in runners
A common injury, and a figure that is easy to misread.
A systematic review from 2021 collected studies of running injuries. In the prospective studies, a mean of 40% of runners were injured during follow-up. The spread between studies was wide, with a standard deviation of 19 percentage points. 9
Achilles tendinopathy, long-lasting pain and reduced function in the tendon, was the most common diagnosis among new injuries. It accounted for 10.3% of them. That is a share of injuries, not of runners. If 40 of 100 runners are injured and one in ten injuries is in the Achilles tendon, about 4 of the 100 runners have Achilles tendinopathy, not 10.
Data · Running-related injuries
The five most common diagnoses, as a share of all injuries.
New injuries: Achilles tendinopathy 10.3%, medial tibial stress syndrome 9.4%, patellofemoral pain 6.3%, plantar fasciitis 6.1%, ankle sprain 5.8%.
Some groups are affected far more often. A Finnish study asked 785 former elite male athletes and 416 matched controls in old age about tendon problems diagnosed by a doctor. Before the age of 45, Achilles tendinopathy had been far more common in former middle- and long-distance runners than in the controls, with an adjusted odds ratio of 31.2. Ruptures of the tendon were much more common in former sprinters than in the controls, with an odds ratio of 14.9. 10 The answers were given decades later, so some diagnoses may have been forgotten or misremembered.
Treatment
Load is part of the treatment.
If loading changes the tendon, loading is a natural place to start when it hurts. Two randomised trials from Scandinavia tested this directly.
- Who took part
- 58 adults with Achilles tendinopathy in the middle of the tendon for more than 3 months, in a Danish trial by Beyer and colleagues. 11
- What was compared
- 12 weeks of eccentric training, where the calf lowers the heel slowly under load, against heavy slow resistance training, where the calf lifts and lowers heavy loads slowly in a gym.
- What they found
- Both groups improved in symptoms and function, and the gains lasted to 52 weeks. The tendon became thinner and had fewer new blood vessels in both groups. The groups did not differ in any of these. Training sessions were completed more often in the heavy slow group, 92% against 78%.
- What it cannot say
- How either programme compares with no exercise, because there was no group without training.
- Who took part
- 38 adults with Achilles tendinopathy in a Swedish trial by Silbernagel and colleagues. 12
- What was compared
- Both groups did the same tendon-loading programme. One group kept running and jumping, guided by a model that set limits for how much pain was acceptable. The other group stopped running and jumping for the first 6 weeks.
- What they found
- Both groups improved by about the same amount. On the 100-point VISA-A score of symptoms and function, the group that kept running went from 57 to 85 at 12 months, and the group that rested went from 57 to 91. The difference in improvement was not significant.
- What it cannot say
- Whether continued running is safe without limits on pain. The running was guided throughout, and 19 people per group is a small trial.
A living review from 2021 pooled 29 randomised trials of 42 different treatments. For tendinopathy in the middle of the tendon, all active treatments seemed better than waiting and seeing at 3 months, and no active treatment was clearly better than the others. The certainty was very low to low, and no trial had a low risk of bias. The authors concluded that clinicians could consider starting with a calf exercise programme because it is cheap, easy to prescribe and has few harms. 13
What the evidence supports
Progressive calf loading as a reasonable first treatment for tendinopathy in the middle of the tendon, with more than one way of doing it.
What it does not support
One exercise as the only right one, complete rest as the default, or a fixed timeline that fits everyone.
Limits of this explanation
Some signs need an assessment, not an article.
Urgent advice
A snap or popping noise at the time of injury, severe pain, a foot or ankle that has changed shape, or not being able to walk. The NHS advises getting urgent advice through its 111 service. 14 A sudden snap at the back of the heel can be a rupture of the tendon.
A planned assessment
Pain that stops normal activities, gets worse or keeps coming back, or has not improved after two weeks of looking after it at home. Tingling or loss of feeling in the foot also belongs here. 14
Diabetes
Foot problems can be more serious with diabetes, and foot pain should be checked by a doctor. 14
The examples are not complete and they are not a triage tool.
What this article cannot tell you
The anatomy comes from dissections of body donors, most of them old, and the twist was measured in tissue that no longer carried load. The training studies measured healthy tendons for a few months. The injury figures are averages across studies with different definitions. None of it can say what is happening in one person's heel. That needs an examination.
Sources
Follow the evidence.
- Gray H. Anatomy of the Human Body. 20th edition, revised by Lewis WH. Philadelphia: Lea and Febiger; 1918. The muscles and fasciae of the leg, Fig. 438. Fig. 438 on Wikimedia Commons Supports: the plate traced in Model 1, the origins of the gastrocnemius, soleus and plantaris, and the insertion on the calcaneus.
- Komi PV, Fukashiro S, Järvinen M. Biomechanical loading of Achilles tendon during normal locomotion. Clinics in Sports Medicine. 1992;11(3):521–531. PubMed record Supports: direct measurement of tendon force in volunteers, high forces in hopping, and the suggestion that the rate of loading matters clinically.
- Edama M, Kubo M, Onishi H, et al. The twisted structure of the human Achilles tendon. Scandinavian Journal of Medicine and Science in Sports. 2015;25(5):e497–e503. PubMed record Supports: three twisted fascicle bundles, the nearly parallel course of the medial gastrocnemius, and soleus torsion 3–5 cm above the insertion, in 111 legs.
- Pękala PA, Henry BM, Ochała A, et al. The twisted structure of the Achilles tendon unraveled: a detailed quantitative and qualitative anatomical investigation. Scandinavian Journal of Medicine and Science in Sports. 2017;27(12):1705–1715. PubMed record Supports: the mean rotation angles (28°, 136°, 129°), the insertion shares (44%, 28%, 28%), the free tendon length of 60.8 mm and the smaller cross-section at the muscle end, in 106 legs.
- Edama M, Kubo M, Onishi H, et al. Structure of the Achilles tendon at the insertion on the calcaneal tuberosity. Journal of Anatomy. 2016;229(5):610–614. PubMed record Supports: the three twist types defined by the deep layer at the insertion, and their frequencies of 24%, 67% and 9% in 132 legs.
- Bohm S, Mersmann F, Arampatzis A. Human tendon adaptation in response to mechanical loading: a systematic review and meta-analysis of exercise intervention studies on healthy adults. Sports Medicine Open. 2015;1(1):7. PubMed record Supports: 27 studies, 264 participants, the pooled changes in stiffness, modulus and cross-sectional area, and larger effects with high loads and 12 weeks or more.
- Lazarczuk SL, Maniar N, Opar DA, et al. Mechanical, material and morphological adaptations of healthy lower limb tendons to mechanical loading: a systematic review and meta-analysis. Sports Medicine. 2022;52(10):2405–2429. PubMed record Supports: 61 studies, 763 participants, the pooled changes, and modulus as the main driver of greater stiffness.
- Heinemeier KM, Schjerling P, Heinemeier J, Magnusson SP, Kjaer M. Lack of tissue renewal in human adult Achilles tendon is revealed by nuclear bomb 14C. The FASEB Journal. 2013;27(5):2074–2079. PubMed record Supports: the bomb-pulse method, 28 tendon and 4 muscle samples, and a tendon core that reflects the first 17 years of life.
- Kakouris N, Yener N, Fong DTP. A systematic review of running-related musculoskeletal injuries in runners. Journal of Sport and Health Science. 2021;10(5):513–522. PubMed record Supports: overall incidence of 40.2% and the shares of injuries by diagnosis.
- Kujala UM, Sarna S, Kaprio J. Cumulative incidence of Achilles tendon rupture and tendinopathy in male former elite athletes. Clinical Journal of Sport Medicine. 2005;15(3):133–135. PubMed record Supports: the odds ratios of 31.2 for tendinopathy in distance runners and 14.9 for rupture in sprinters.
- Beyer R, Kongsgaard M, Hougs Kjær B, et al. Heavy slow resistance versus eccentric training as treatment for Achilles tendinopathy: a randomized controlled trial. American Journal of Sports Medicine. 2015;43(7):1704–1711. PubMed record Supports: equal and lasting improvement with both programmes and the difference in completed sessions.
- Silbernagel KG, Thomeé R, Eriksson BI, Karlsson J. Continued sports activity, using a pain-monitoring model, during rehabilitation in patients with Achilles tendinopathy: a randomized controlled study. American Journal of Sports Medicine. 2007;35(6):897–906. PubMed record Supports: no negative effect of guided continued running and jumping, and the VISA-A scores in both groups.
- van der Vlist AC, Winters M, Weir A, et al. Which treatment is most effective for patients with Achilles tendinopathy? A living systematic review with network meta-analysis of 29 randomised controlled trials. British Journal of Sports Medicine. 2021;55(5):249–256. PubMed record Supports: active treatments ahead of wait-and-see at 3 months, no clear difference between them, and the very low to low certainty.
- NHS. Ankle pain. NHS page on ankle pain Supports: the signs that call for urgent advice or a GP appointment.
Continue
Where to go from here.
All Insights · Pain and context · How to read evidence
General education, not a diagnosis or an individual treatment plan. Models are schematic unless a data source is stated beside the figure. Zacharias Razvi is a final-semester physiotherapy student and is not yet an authorised physiotherapist.